ASHRAE TC 2.6 PRESENTSORLANDO 2005 WHAT DID WE LEARN FROM ASHRAE RP-879?

Transcription

1 ASHRAE TC 2.6 PRESENTSORLANDO 2005 WHAT DID WE LEARN FROM ASHRAE RP-879? Norm Broner Operations Manager Vipac Engineers and Scientists Ltd, Australia

2 RP-879 ASHRAE sponsored research on the Relationship between LF HVAC noise and Comfort in occupied Spaces Phase 1 showed that three factors important: overall level spectral imbalance temporal modulation Phase 2 involved psycho-acoustic testing to derive metrics and criteria

3 Acoustical (Sound) Quality We want sound that is not intrusive The optimum sound quality is one that occurs when a sound is not particularly noticeable with respect to our on-going activities - the sound is neutral The sound should have certain spectral and temporal characteristics.

4 Neutral Spectrum?

5 Annoyance vs Spectrum Imbalance & Level

6 Noise Stimuli - Ten base stimuli were drawn from the noise recordings of over 70 HVAC noises that were collected during Phase 1 - two samples were neutral (spectra with no specific spectral characteristics and with slopes close to -5dB/octave) - two Neutral/marginal (spectra with slight deviations from neutral), - three Rumble (spectra with some degree of rumble) and - three Strong Rumble (spectra with significant low frequency energy relative to neutral).

7 Noise Stimuli - A total of 60 HVAC noise stimuli were created from the base samples - For the purpose of testing, the 60 stimuli were further randomly divided into groups of 15, thus generating four Sequences designated A, B, C and D - The only proviso was that more or less equal numbers of each type of stimulus should occur in each of the four sequences - The nominal overall Sound Pressure Levels ranged from db and dba respectively

8 Strong Rumble Noise Stimulus

9 Neutral Noise Stimulus

10 Noise Stimuli Playback SOUND FILES SC-63 PARAMETRIC EQUALISER JBL MPX600 TEST ROOM

11 Response Rating The method of Magnitude Estimation was used. To each sound, assign a number that corresponds to the annoyance. You may use any positive number that appears appropriate to you - whole numbers, decimals or fractions. Do not worry about running out of numbers, there will always be a smaller number than the smallest you use and a larger one than the largest you use. Simply try to match an appropriate number to each noise regardless of what you may have called some previous sound.

12 Response Rating For each of the sounds, record the annoyance defined as the nuisance aspect of the sound experienced. Imagine you are in an office and are seated in your chair while working. Please estimate how annoyed you would feel when exposed to each sound. For each of the sounds, record its loudness as defined as the perceptual aspect of the noise that is changed by turning the volume knob on a radio or TV.

13 Sample Rating Form Name: KB Age: 55 Sex: M Do you regard yourself as being sensitive to noise? Yes Y SEQ A Stimulus Rating Rating Relief Acceptable Comm No. Loudness Annoyance Yes No Yes No * * * * * * * * * * * * * * * * * * * * * * * *

14 Test Room Double wall construction and floated on isolators 6700 long x 3100 wide x 2350 high Noise stimuli played back through two loudspeakers

15 Test Room

16 Test Protocol Gender of Subjects Age of Subjects Other Information White Noise Test Final Main Study 21 Subjects : 11 male 10 female 19 55: Average years Only two reported not sensitive to noise Each subject rated 9 stimuli: dba - 21 x 9 ratings used in analysis Each subject rated 4 x 45 stimuli - 21 x 4 x 2 x 15 ratings used in analysis -Relief and Acceptability also rated

17 Annoyance vs Loudness Annoyance vs Loudness for 60 Main Study Stimuli Annoyance Rating y = x R 2 = Loudness Rating

18 A/L Ratio Results Annoyance/Loudness Ratio vs SPL (dba) for 60 Main Study Noise Stimuli 2.0 A/L Ratio y = x R 2 = A/L A/L = 1.0 Linear (A/L) SPL (dba)

19 Example Regression Models R 2 =0.87 Model 1: Nmn R 2 =0.92 Model 2: Nmn Amax R 2 =0.93 Model 3: Nmn Amax S R 2 =0.93 Model 4: Nmn Amax S CohAC R 2 =0.80 Model 5: PA Predicted Response of Subjects Measured Response of Subjects

20 Results Simple regression models gave reasonable results (R 2 = ) Regression (Cmax A) result poor Regression using Cmax and A R 2 = 0.87

21 Review of Results - Metrics Fluctuations at low frequencies are important for annoyance as the more annoying samples had a large fluctuation in C but only a small variation in A It seems that in these tests (10 seconds duration only), Loudness was just too dominant and that the influence of other variables on annoyance did not have time to develop fully

22 Review of Results - Metrics The data suggests that the adaptation rates for Loudness and Annoyance are different at low sensation levels and that they diverge with time What is needed is to combine some measure of Loudness with measures of spectral balance and signal level fluctuation and also a measure that takes into account the rate-of-change of signal level

23 Criteria for HVAC Are there any design guidelines in terms of Chapter 47 Sound and Vibration?

24 Criteria for HVAC Sequence A Stimulus No. Based On SPL db dba RC QAI SubL SubA A/L Ratio 1 R# RC 31(LFVb) N# RC 30(LF) SR#41b RC 21(LFVa) R# RC 30(LFVa) N# RC 26(LF) N# RC 35(MF) SR# RC 33(LFVb) R# RC 20(LFVb) NM# RC 30(LFVa) SR# RC 32(LFVb) NM# RC 36(LFVa) N# RC 22(LF) SR# RC 31(LFVa) N# RC 29(LFVb) R# RC 35(LFVa) Sequence B

25 Criteria for HVAC A review of the subjective ratings including A/L ratios, RC and QAI values shows: When the RC value is low (RC < 25) and QAI high (> 25), A/L ratio is increased A/L Ratios above 1.2 occur when QAI > 25 Thus, one cannot look at the absolute RC value alone. One also has to look at the Sound Quality in terms of the QAI

26 Criteria for HVAC Sig N Based On SPL db dba RC QAI A/L Ratio 30 R# RC 21(LFVa) SR#41b RC 25(LFVa) R# RC 21(LFVa) SR# RC 21(LFVa) SR#41b RC 24(LFVa) SR# RC 32(LFVa) SR# RC 24(LFVa) SR# RC 25(LFVa) SR# RC 32(LFVb) SR# RC 18(LFVb) R# RC 35(LFVa) SR# RC 34(LFVa) SR# RC 24(LFVa) SR# RC 21(LFVa) NM# RC 30(LFV ) Rank ordering the 60 stimuli showed: the highest A/L ratios were for stimuli based on strong rumble and rumble the lowest A/L ratios for stimuli with low RC and QAI

27 Criteria for HVAC A/L Ratios vs RC for the 60 Main Study Stimuli A/L Ratio y = x R 2 = A/L Ratio A/L = 1.0 Linear (A/L Ratio) RC A/L Ratios vs QAI for 60 Main Study Stimuli A/L Ratio y = x R 2 = A/L Ratio A/L = 1.0 Linear (A/L Ratio) QAI

28 Criteria for HVAC Loudness and Annoyance for RC vs QAI Subjective Rating y = x R 2 = y = x R 2 = y = x R 2 = A/L Ratio SubL SubA A/L Ratio Linear (SubA) Linear (SubL) Linear (A/L Ratio) QAI Loudness and Annoyance for RC vs QAI Subjective Rating y = x R 2 = y = x R 2 = y = x R 2 = A/L Ratio SubL SubA A/L Ratio Linear (A/L Ratio) Linear (SubL) Linear (SubA) QAI 0.8

29 Criteria for HVAC At low RC (< 30), a QAI > 25 is required to generate a significant A/L As the RC value increases, the increase loudness associated with higher RC levels can induce annoyance even if the QAI is not large For RC 30 35, with a QAI in the range 25 30, the HVAC sound would be considered at least annoying, if not strongly annoying

30 Criteria for HVAC Recommended Criteria in terms of RC and QAI Subjective Response RC QAI Neutral <30 <25 Acceptable <10 Annoying Strongly Annoying >35 > 30

31 Summary - Criteria for HVAC HVAC criteria based on RC values alone are not adequate. Sound Quality factors also need to considered In terms of criteria, RC > 30 can lead to excessive annoyance when the spectral balance is significantly different form neutral (QAI > 25 30)

32 Conclusions To improve the low frequency noise assessment capability it is necessary to combine some measure of Loudness with a measure of spectral balance and signal level fluctuation and also a measure that takes into account the rate of change of the signal level (i.e. the time period over which the level fluctuation is occurring Future research should focus on testing with a minimum of a one-hour exposure to noise stimuli and the subjects should conduct some form of performance or vigilance task

33 Other Conclusions Strong implication that for HVAC noises with rumble and temporal variations, Loudness and Annoyance are NOT the same The difference between Loudness and Annoyance increases with decreasing frequency and with time This research supports the RC method of HVAC assessment This research suggests that the NCB curves below 100 Hz may need to be revised